GitHub

Welcome to Cody! An AI assistant designed to let you interactively query your codebase using natural language. By utilizing vector embeddings, chunking, and OpenAI's language models, Cody can help you navigate through your code in an efficient and intuitive manner. 💻
https://www.star-history.com/#ajhous44/cody&Date
LINK
https://www.loom.com/share/eba1d0dcee20430fbd412580d1c0ea0e?sid=4998cf6f-45b4-480d-b742-6f22f3a49dc3
Cody continuously updates its knowledge base every time you save a file, ensuring you have the most up-to-date information. You can customize your setup by specifying directories to ignore in the ignore_list.
🚀 Getting Started
1. Clone the repo
2. (Optionally) Setup virtual environment by running pip install -m venv .venv and then pip install -r requirements.txt in terminal from the root of your directory
3. Rename the .local.env file to .env and replace YOUR_API_KEY_HERE with your OpenAI API Key.
4. Modify the IGNORE_THESE global var at the top of the script to specify directories and files you wish to exclude from monitoring. (You should comment out any large files like a virtual environment, cache, js libraries you have downloaded, etc...)
5. Run the script using Python: python cody.py and follow terminal for setup. It will prompt you for if you want to use text chat (terminal) or conversational (speech i/o). It will also warn you if you remove .env from the ignore list.
🎯 Features
- **File Monitoring**: Real-time monitoring of all files in your project's directory and subdirectories. 👀
- **Embedding-based Knowledge Base**: Create a knowledge base using OpenAI Embeddings. Cody collects the contents of all text and JSON files and adds them to this knowledge base. 📚
- **Interactive Q&A**: Listen to user inputs. Ask questions, and Cody will generate a response using the knowledge base. 🧠
- **Customizable**: Easily specify files or directories to ignore during monitoring.
🛠 Dependencies
- dotenv: Load variables from a .env file into the environment.
- langchain-community: A language processing library used for embeddings and vector storage. Previously langchain.
- langchain_openai: Provides the OpenAIEmbeddings functionality, integrating OpenAI models directly with langchain's architecture.
- litellm: Call all LLM APIs using the OpenAI format (https://github.com/BerriAI/litellm)
- watchdog: Monitor filesystem events in real-time.
- openai: Generate smart responses using OpenAI's language model.
- speech_recognition: Convert speech to text for voice interaction.
- gtts: Google Text-to-Speech library for generating audio from text.
- pygame: Library to play audio files.
💡 Usage
- To stop the script, type 'exit' or speak the word 'exit' and press enter. Cody will gracefully terminate the program.
Configuring the Ignore List
Cody allows you to specify which files and directories should be ignored during file monitoring. This is particularly useful for excluding files that change frequently, are not relevant to your queries, or could contain sensitive information.
To customize your ignore_list, add patterns matching the files or directories you wish to exclude. Cody supports simple wildcard patterns for flexibility. Here are some examples to guide you: